Dimension expansion and customized spring potentials for sensor localization

EURASIP Journal on Advances in Signal Processing

Table 1 Comparison of average convergence rates under different noise levels

Noise level	Spring type	Error rate at iterations of multiples of 100
σ = 0.00	Soft spring	0.2467	0.1468	0.1514	0.1402	0.0702	0.0419	0.0256	0.0174
	Strong spring	0.2501	0.1568	0.1796	0.1077	0.0750	0.0437	0.0277	0.0189
	“Lock-in” mode	0.1347	0.1651	0.1013	0.0527	0.0325	0.0213	0.0144	0.0100
σ = 0.05	Soft spring	0.2486	0.1458	0.1499	0.1406	0.0693	0.0419	0.0259	0.0181
	Strong spring	0.2508	0.1566	0.1784	0.1085	0.0745	0.0437	0.0285	0.0190
	“Lock-in” mode	0.1340	0.1651	0.1004	0.0519	0.0323	0.0216	0.0151	0.0128
σ = 0.10	Soft spring	0.2485	0.1478	0.1528	0.1389	0.0710	0.0427	0.0270	0.0204
	Strong spring	0.2507	0.1569	0.1791	0.1082	0.0742	0.0438	0.0288	0.0215
	“Lock-in” mode	0.1363	0.1640	0.1016	0.0528	0.0334	0.0236	0.0199	0.0214

Table1 records the average error rate of 100 independent simulations, at iterations of multiples of 100. The results in Table1 demonstrate that the algorithm with customized spring potential works consistently better than the other two cases.